The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to analyze external C source code files for vulnerabilities, which introduces a potential indirect prompt injection surface where malicious code comments could attempt to influence the agent's behavior. * Ingestion points: C source code files (SKILL.md). * Boundary markers: The instructions do not define explicit delimiters or instructions to ignore embedded commands within the analyzed code. * Capability inventory: The skill explicitly directs the agent to execute shell-level commands (radare2 oneliners) based on findings. * Sanitization: No sanitization process is defined for the agent's ingestion of the target source code.\n- [DYNAMIC_EXECUTION]: The skill recommends that the agent dynamically generate and execute 'radare2 oneliners' to verify if identified vulnerabilities are real. This represents a dynamic execution pattern that, while central to the skill's auditing purpose, involves executing commands that could be influenced by the untrusted data being analyzed.

aibugs